Appendix 1. Details of the mobbing call recording in the playback experiments.
Objectives
Although it is desired to use different recording in each playback experiment (Kroodsma et al., 2001), we used a single mobbing call recording recorded in the breeding season in all playback experiments. To justify this, we assessed whether the recording can be considered a representative of mobbing calls of our focal species.
Specifically, using many available mobbing call recordings in xeno-canto
(http://www.xeno-canto.org/, accessed at 0130-2017), we compared the number of discrete mobbing calls of our focal species in our recording with those in other regions because the Paridae species has several types of mobbing calls that might convey different information (Hailman, 1989). In addition, we examined whether the number of discrete mobbing calls differs between breeding and non-breeding seasons. In these assessments, we focused on one of our focal species (the marsh tit) for following reasons: 1) it is known that various species including the family Paridae respond to not only conspecific mobbing calls but also interspecific ones (e.g., Bélisle and Desrochers, 2002). 2) The marsh tit has wide distribution range across the old world as well as is one of the most common Paridae species in our study region. The following analyses were performed using Adobe Audition 3.0 software and R software (ver. 3.2.0) with “tuneR”
and “seewave” packages.
Compiling mobbing call recordings from xeno-canto
In available recordings of the marsh tit in xeno-canto, we compiled 43 mobbing call recordings which were originally categorized as “alarm call” or “aggression call”.
Of these, we excluded recordings with high levels of background noise (e.g., wind noise, other bird calls) from further analyses. We also excluded recordings that accompanied with marsh tits’ songs because “alarm calls” uttered in such situation might convey different information compared to those uttered to predators. As a result, we used 30 mobbing call recordings (mean recording times (sec) ± Standard Deviation
= 60.40 ± 39.78) recorded in several countries (14 in Germany, 9 in Poland, 3 in France,
2 in Netherlands, 1 in Czech and Sweden) in the following analyses.
Comparisons of the number of discrete mobbing calls in different locations and seasons
Based on the sound spectrograms, we counted the number of spectrally discrete mobbing calls for our single recording, all xeno-canto recordings (hereafter the xeno- canto recordings, n = 30), and subsets of the xeno-canto recordings in the breeding season (March to August, hereafter the nonbreeding recordings, n = 8) and in the non- breeding season (September to February, hereafter the nonbreeding recordings, n = 22).
We identified four spectrally different types of mobbing calls of the marsh tit in all recordings we assessed (Fig A1.1). Hereafter, we referred to each type of mobbing calls as type I, II, III, and IV calls, respectively. Among them, our 30-sec recording included three types of mobbing calls (types I, II and III). Specifically, the recording included 20 type I calls, 13 type II calls and 6 type III calls. The type I calls consisted of a single “tiii” note (Fig A1.1a). The type II calls consisted of a single “chicka” note and various number of “dee” notes (Fig A1.1b), which is an important metric of Paridae’s mobbing call quality (Courter and Ritchison, 2010). The number of “dee” notes in type II calls (mean ± SD) was 2.31 ± 1.18 (Range: 0-5). The type III calls consisted of multiple “jee” notes (Fig A1.1c) and the number of “jee” notes (mean ± SD) was 9.50 ± 3.27 (Range: 5-13). The xeno-canto recordings, the nonbreeding recordings and the breeding recordings had four different types of mobbing calls (types I, II, III and IV).
The mean numbers of different types of mobbing calls (± SD) were 1.30 ± 0.47 in the xeno-canto recordings (n = 30), 1.26 ± 0.49 in the breeding recordings (n = 8) and 1.30
± 0.47 in the nonbreeding recordings (n = 22). The seasonal difference (breeding vs.
non-breeding seasons) was not significant (t-test; p = 0.77). The type IV calls consisted of multiple “pi” or “picho” notes (Fig A1.1d) and were included in only two recordings, suggesting that this type of mobbing call would be rare. Thus, these results suggest that our single recording includes almost all types of mobbing calls of the marsh tit and there are no seasonal variations of their mobbing calls, indicating that our 30-sec recording can be considered a representative mobbing calls of our focal species.
References
Bélisle, M., and A. Desrochers. 2002. Gap-crossing decisions by forest birds: an empirical basis for parametrizing spatially-explicit, individual-based model.
Landscape Ecology 17:219-231.
Courter, J. R., and G. Ritchison. 2010. Alarm calls of tufted titmice convey information about predator size and threat. Behavioral Ecology 21:936-942.
Hailman, J. P. 1989. The organization of major vocalizations in the Paridae. Wilson Bulletin 101:305-343.
Kroodsma, D. E., B. E. Byers, E. Goodale, S. Johnson, and W.-C. Liu. 2001.
Pseudoreplication in playback experiments, revisited a decade later. Animal Behaviour 61:1029-1033.
Fig. A1.1. Examples of sound spectrograms of the mobbing calls of the marsh tit. We draw spectrograms of the type I, II, III calls using the recording used in the playback
experiments. The sound spectrogram of the type IV call was drawn using one of xeno- canto recordings (XC127377).